[Code of Federal Regulations]
[Title 40, Volume 7]
[Revised as of July 1, 2005]
From the U.S. Government Printing Office via GPO Access
[CITE: 40CFR60.C]
[Page 674]
Appendix C to Part 60--Determination of Emission Rate Change
1. Introduction
1.1 The following method shall be used to determine whether a
physical or operational change to an existing facility resulted in an
increase in the emission rate to the atmosphere. The method used is the
Student's t test, commonly used to make inferences from small samples.
2. Data
2.1 Each emission test shall consist of n runs (usually three) which
produce n emission rates. Thus two sets of emission rates are generated,
one before and one after the change, the two sets being of equal size.
2.2 When using manual emission tests, except as provided in Sec.
60.8(b) of this part, the reference methods of appendix A to this part
shall be used in accordance with the procedures specified in the
applicable subpart both before and after the change to obtain the data.
2.3 When using continuous monitors, the facility shall be operated
as if a manual emission test were being performed. Valid data using the
averaging time which would be required if a manual emission test were
being conducted shall be used.
3. Procedure
3.1 Subscripts a and b denote prechange and postchange respectively.
3.2 Calculate the arithmetic mean emission rate, E, for each set of
data using Equation 1.
[GRAPHIC] [TIFF OMITTED] TC01JN92.292
Where:
Ei=Emission rate for the i th run.
n=number of runs.
3.3 Calculate the sample variance, S2, for each set of data using
Equation 2.
[GRAPHIC] [TIFF OMITTED] TC01JN92.293
3.4 Calculate the pooled estimate, Sp, using Equation 3.
[GRAPHIC] [TIFF OMITTED] TC01JN92.294
3.5 Calculate the test statistic, t, using Equation 4.
[GRAPHIC] [TIFF OMITTED] TC01JN92.295
4. Results
4.1 If Eb,Ea and tt', where t' is the critical
value of t obtained from Table 1, then with 95% confidence the
difference between Eb and Ea is significant, and an increase in emission
rate to the atmosphere has occurred.
Table 1
------------------------------------------------------------------------
t' (95
percent
Degrees of freedom (na=nb-2) confidence
level)
------------------------------------------------------------------------
2.......................................................... 2.920
3.......................................................... 2.353
4.......................................................... 2.132
5.......................................................... 2.015
6.......................................................... 1.943
7.......................................................... 1.895
8.......................................................... 1.860
------------------------------------------------------------------------
For greater than 8 degrees of freedom, see any standard statistical
handbook or text.
5.1 Assume the two performance tests produced the following set of
data:
------------------------------------------------------------------------
Test a Test b
------------------------------------------------------------------------
Run 1. 100.................................................. 115
Run 2. 95................................................... 120
Run 3. 110.................................................. 125
------------------------------------------------------------------------
5.2 Using Equation 1--
Ea=100+95+110/3=102
Eb=115+120+125/3=120
5.3 Using Equation 2--
Sa2=(100-102) 2+(95-102) 2+(110-102) 2/
3-1=58.5
Sb2=(115-120) 2+(120-120) 2+(125-120)
2/3-1=25
5.4 Using Equation 3--
Sp=[(3-1)(58.5)+(3+1)(25)/3+3-2] \1/2\=6.46
5.5 Using Equation 4--
[GRAPHIC] [TIFF OMITTED] TC01JN92.296
5.6 Since (n\1\+n\2\-2)=4, t'=2.132 (from Table 1). Thus since t>t'
the difference in the values of Ea and Eb is significant, and there has
been an increase in emission rate to the atmosphere.
6. Continuous Monitoring Data
6.1 Hourly averages from continuous monitoring devices, where
available, should be used as data points and the above procedure
followed.
[40 FR 58420, Dec. 16, 1975]